Alva uses AI to help you (a candidate) save time in applying for jobs and to provide you with relevant job recommendations based on your previous experience and interests. Our AI-powered features are optional to use and are continuously reviewed by humans to ensure accuracy, relevance, and consistency.
Using AI to Extract CV Data & Match You with Jobs
After completing Alva’s tests, you have the option to save and reuse your test results to find other job opportunities that match your profile. You may provide additional information to increase the relevance of jobs recommended to you by manually uploading information, uploading your CV, or connecting your LinkedIn account.
If you choose to populate your Marketplace profile using your LinkedIn account, Alva uses LinkedIn’s *Application Programming Interface (API) to fetch data from your LinkedIn profile. If you opt to upload your CV to Alva’s Marketplace platform instead of manually entering the data or using LinkedIn, Alva uses Gemini 1.5 Pro (a large language model, or “LLM”) to extract data on your work experience, language, and location. You will then be asked to verify the extracted data and will be asked to enter any missing information manually.
After your data is compiled, it is sent through an embedding model to determine whether your background and skills match the requirements of job positions available in Alva’s Marketplace. If your background and interests align with the description of a job position, you will receive a recommendation for that role.
How Does The Embedding Model Work?
The embedding model is a crucial part of how Alva's Marketplace connects candidates with job opportunities. Here's a simple breakdown of how it functions:
Data collection: We start by gathering information about you from your LinkedIn profile, CV, or any details you enter manually. This includes details about your work experience, skills, and location.
Standardization: This information is then transformed into a standardized format, making it easier for our system to process and understand.
Creating vectors: Using the embedding model, your data and the data from job descriptions are converted into "vectors”, or numbers that represent the data from the text. Think of vectors as unique fingerprints for each candidate and job, capturing the essence of the text in a way that a computer can analyze.
Storing vectors: These vectors are stored securely in our system. Each candidate and job has its own vector, which helps computers compare your skills and backgrounds to the job positions available in Alva Marketplace.
Matching process: Our system calculates the "distance" between your vector and the vectors of all available jobs. Jobs that are a close match (meaning their vectors are similar to yours) are shown to you as recommendations on your recommendations page.
By using this embedding model, we can ensure that the job recommendations you receive are not only relevant but also tailored to your unique background and skills. This process allows us to present you with opportunities that best fit your profile, even if you might not have initially considered them.
Classifying Job Positions
In addition to the extraction and matching processes, Alva uses AI to sort jobs in our Marketplace into categories. For example, an “Account Executive” position would be categorized as “Sales”. This allows Alva to provide you with recommendations for roles that are within the job families you are interested in. Recruiters posting job positions in Marketplace have the opportunity to check if the suggested classification is accurate and to correct this if needed.
What Does Alva Do To Ensure Safe AI?
Your Data, Your Rights
You can reach out to Alva’s customer support team at any time to revoke Alva’s access to your data or to correct it. If you decide to connect your LinkedIn account, Alva will have access to your LinkedIn data for one year. Your account data will be retained for two years unless you request for the data to be deleted. You can also revoke Alva’s access to your LinkedIn account through LinkedIn’s settings at any time.
Safe AI
Alva takes proactive measures to ensure that our use of AI and ML not only meets but exceeds industry standards for data protection and ethical practice. Our GDPR compliance demonstrates our dedication to protecting user data and privacy, and our certifications in ISO 27001 and DNV (for both the logic and personality tests) reflect our robust approach to information security and responsible AI practices. Alva has also established an AI Management System based on the ISO 42001 framework and continuously remediates risks around AI to ensure our product and operations sufficiently protect the interests of individuals and society.
These commitments enable us to responsibly harness the power of AI and ML, providing our customers with the assurance that their data is handled with the utmost care and integrity.
Questions or Concerns?
If you have any questions about our use of AI or feel that you have been negatively impacted by Alva’s use of AI, please reach out to support@alvalabs.io or ai@alvalabs.io.